From AI Tool to AI Co‑Worker: What Dassault Has Actually Launched
Dassault Systèmes’ AI virtual companions AURA, LEO and MARIE are embedded agents on the 3DEXPERIENCE platform that interpret user intent, apply industry knowledge and execute design, simulation, manufacturing and operations tasks across 19 defined competencies to support end‑to‑end industrial workflows in a way that aims to turn AI from a separate tool into a day‑to‑day co‑worker within enterprise processes. This is not another chat window bolted onto CAD. Dassault has announced these virtual companions for program management (AURA), complex engineering (LEO) and deep science (MARIE) directly inside its cloud‑based platform, alongside a vision for “virtual twin factories.” By framing the agents as workers, the company is signaling a shift in manufacturing workflows: AI is meant to participate in the work itself, not merely comment on it. That is the core takeaway for any team wondering if these agents matter beyond the marketing language.
Nineteen Competencies: Where AI Touches the Workflow
The most important detail is not the three names but the 19 competencies they cover. Dassault says AURA, LEO and MARIE “support workflows across design, simulation, manufacturing and operations with 19 competencies,” a clear statement that AI design automation is now spread across the whole lifecycle, not confined to single tasks. Each competency is an area of knowledge—like mechanical design—made of concrete skills, such as conceptual sketching or detailed modeling. These competencies guide users, reveal invisible insights, explore untapped spaces and execute relevant tasks across business, product design, engineering, simulation, manufacturing and operations workflows. In plain terms, the agents are being trained to handle units of work, not vague conversations: they interpret intent, reason in an industrial context, make decisions and generate outcomes grounded in science and industry‑accurate reality. That is the promise users will judge in production environments.
AURA, LEO and MARIE: Three Personalities, One Strategic Direction
Dassault deliberately gives each virtual companion a role and a temperament, which says a lot about how it expects AI to fit into teams. Aura is described as agreeable and meant for program management, Leo is assertive and made for engineering, and Marie—focused on deep science—is even more assertive. This character design is more than branding; it reflects a view that AI agents will take on different levels of initiative depending on the task. Program management needs a cooperative coordinator, engineering benefits from a confident problem‑solver and science often demands a forceful challenger. More importantly, all three AI virtual companions are live on the 3DEXPERIENCE platform, not as external bots but as embedded workers operating on AI factories from OUTSCALE, Dassault’s cloud and AI operator. The infrastructure is pitched as industrial‑grade and protective of customer data and IP, which will be a major adoption barrier if it fails and a strategic edge if it holds.
From Virtual Twins to Virtual Twin Factories
Dassault has long talked about virtual twins of products and plants; now it is promoting “virtual twin factories” that include AI companions as workers on par with humans. In this vision, the factory of the future brings together virtual companions as digital workers, generative experiences as highly automated processes and “virtual twin as a service” as predefined industry world models. That is an ambitious expansion: it tries to make the 3DEXPERIENCE platform an AI‑native environment where simulation models, manufacturing workflows and operations analytics are all run by coordinated agents. The practical impact for ordinary users could be significant if it works. Instead of manually stitching together tools, teams could ask AURA to replan a program, let LEO refine and validate a design in simulation, then have MARIE interrogate deep scientific assumptions—all within the same virtual twin factory context. That is a far cry from today’s fragmented enterprise stacks.
What This Means for Engineers Right Now—and What Comes Next
For engineers and operations teams, the immediate impact is the arrival of embedded AI agents that act on units of work rather than isolated prompts. Virtual companions use structured industry knowledge and know‑how to interpret user intent, reason in an industrial context, take decisions and produce outcomes aligned with established methods. In day‑to‑day manufacturing workflows, that could mean faster design iterations, more context‑aware simulations and AI design automation that respects process constraints instead of ignoring them. It is fair to remain skeptical until more real‑world deployments are visible; even the coverage notes that more details on virtual twin factories will follow and promises to “report back on all this as more details come in.” Still, Dassault’s move is strategically clear: it is betting that the future of industrial software is not AI on the side, but AI inside—the companion sharing your project timeline, simulation deck and factory model, not a chatbot watching from the browser toolbar.
